> From: bill lam
>
> IMHO if nomenclature in task.ijs is incorrect or causing confusion
> then it is a baggage. It is desirable to get it fixed even that means
> breaking existing code.
It won't bother me too much because I don't have a lot of existing code that
would break. I may feel differently if that was not the case! Perhaps, given
task.ijs is a relatively recent addition to ~system, it is not yet in
widespread use?
The nomenclature changes seem to me to be a separate issue to the proposed code
changes.
Change task.ijs code:
====================
Any objections to these proposed code changes to task.ijs:
* Change 1st line of current definition of fork from
fork=: (3 : 0)`(2!:1)@.IFUNIX
to
fork=: (3 : 0)`([: 2!:1 '(' , ')&' ,~ ])@.IFUNIX
* Change 1st line of current definition of spawn from
spawn=: (3 : 0)`(2!:0...@])@.IFUNIX
to
spawn=: (3 : 0)`([: 2!:0 '(' , ' || true)' ,~ ])@.IFUNIX
Change interface for task.ijs:
==============================
Which proposal for name changes in task.ijs is preferable, A or B?
Current interface
spawn - runs command in host, waits & returns stdout
shell - runs command in shell (cmd for Windows) waits & returns stdout
fork - runs command, doesn't wait, returns ''
launch - opens file with associated application
Proposal A:
(wouldn't break existing code)
Leave "shell", "fork" & "launch" as is.
Create "host" or "hostcmd" as alias for "spawn", deprecate "spawn".
Proposal B:
(would match Dictionary terminology better)
Leave "shell" and "launch" as is.
Rename spawn to "host" or "hostcmd"
Rename "fork" to "spawn".
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
